Animals tend to alternate between different choices, which requires the ability to remember recent choices. The Y-maze spontaneous alternation test is widely used in various animal models for assessing short-term memory, and its precise evaluation depends upon the accurate determination of the arm visit sequence. However, an objective method for defining arm visits is lacking owing to uncertainty regarding the extent to which an animal must go into the arm to be considered visited. Here, we conducted quantitative analyses on mice behavior in the Y-maze while systematically varying the arm visit threshold and assessed the effect of acute social isolation on spatial working memory. Our results revealed that 24-h social isolation significantly reduced spontaneous alternation rate when the arm threshold was set at the distal part of the arm. Furthermore, the memory of the recently visited arms faded away faster in the socially isolated mice. However, other behavioral factors were comparable to those of the group-housed mice, indicating a specific impairment of short-term memory. Our findings suggest that the location of arm visit threshold is critical for the precise evaluation of short-term memory, and our study provides a method for comprehensively and systematically assessing spontaneous alternation behavior in the Y-maze.